a) Industry Background
Petroleum reservoirs, whether primarily oil reservoirs or gas reservoirs, often contain significant quantities of hydrogen sulphide (H2S) and carbon dioxide (CO2) in addition to hydrocarbons. These contaminants must be removed or at least reduced to meet commercial specifications for purity before the natural gas can be marketed to consumers. The hydrogen sulphide and carbon dioxide, usually referred to as “acid gases”, have commercial value as by-products in and of themselves if, for example, the hydrogen sulphide is converted to sulphur and the CO2 is used for miscible flooding of oil reservoirs. Otherwise, the acid gases are considered to have no marketable value, and are disposed of either by pumping down a disposal well or by flaring.
Commercial specifications for natural gas require that essentially all of the hydrogen sulphide be removed from the gas, typically to a final concentration of 4 PPM (parts per million) by volume or less. Carbon dioxide must likewise be reduced, but being non-toxic, the tolerance for CO2 is much higher (typically 2% by volume for commercial pipeline quality gas).
The extremely stringent specification for H2S content in natural gas has dictated the type of process that must be used, and virtually all natural gas being “sweetened” today is treated by one of the various alkanolamines that are available for this purpose. More than half a century ago, the Girbitol process was introduced in which the primary amine, monoethanol amine (popularly known as “MEA”), was used as the absorbent. Since then, other amines have become popular, namely diethanoamine (DEA), and a current favourite, methyldiethanol amine (MDEA), which is popular because of its preferential affinity for hydrogen sulphide over carbon dioxide. In most cases, generic amine in an aqueous solution is used, although various processes are available in which chemical additives are used in the amine solution to enhance certain characteristics of the absorbent. Amine has gained widespread acceptance and popularity because it can produce a natural gas product that reliably meets the strict requirements for gas purity, especially the requirements for hydrogen sulphide, and can do it relatively inexpensively.
Alternative processes for acid gas removal, such as physical absorption in a solvent or distillation for removal of acid gases, have not been used extensively, except possibly for bulk removal followed by cleanup with amine or as a scavenger for small volumes. Amine is able to remove acid gas components by reacting with them, which in an equilibrium situation can potentially totally remove the acidic components from the gas. Acid gases can be removed by other processes based on chemical reaction, such as the hot carbonate process and various forms of the iron oxide process, which can meet the specifications for gas purity. However, for many practical reasons these processes have never gained widespread popularity.
Historically, the primary concern of the gas processing industry has been to produce natural gas that will meet the stringent requirements for gas purity imposed by pipeline and distribution companies who establish the specifications for natural gas. There has been much less attention directed toward the by-product of the amine process; the acid gas mixture of H2S and CO2 that is co-absorbed in the process. Typically, these two gases are not subjected to any separation process to recover them as two separate entities, but are sent directly as feed to a sulphur plant. Most sulphur plants utilize some version of the Claus process in which one third of the H2S is oxidized by combustion to SO2, which then subsequently reacts with the remaining two thirds of the H2S to produce elemental sulphur and water. The second acid gas component, carbon dioxide, is an inert gas and a none-participant in the chemical reaction, but because of the thermodynamics of the Claus process, carbon dioxide will detrimentally affect the reaction to produce sulphur. The presence of carbon dioxide dilutes the reactants—hydrogen sulphide, oxygen, and sulphur dioxide, retarding the reaction and reducing the percentage conversion to sulphur. The dilution effect directly influences the chemical equilibrium of the Claus process, fundamentally reducing the attainment of high rates of sulphur conversion. In cases where the acid gas feed to the sulphur plant is rich in H2S, the effect of dilution by CO2 may not be too serious, but in those cases where the quantity of CO2 exceeds the quantity of H2S by a factor of five or more, the effect on thermodynamic equilibrium conversion to sulphur is very significant.
A secondary effect of dilution of H2S by excessive quantities of CO2 is flame stability in the reaction furnace where H2S is oxidized to SO2. Carbon dioxide is an effective fire extinguishing chemical, and when present in excessive amounts in the reaction furnace it can inhibit combustion, and in some cases completely quench the flame. The dilution effect of CO2 in the firebox of the furnace will also reduce furnace temperature to the extent that complete combustion does not occur. This necessitates the addition of natural gas to the acid gas entering the sulphur plant in order to improve combustion and maintain flame temperature in the reaction furnace. Natural gas in the reaction furnace causes a further complication by increasing the undesirable reaction by-products, carbonyl sulphide and carbon disulphide. These are the products of reaction between methane and other hydrocarbons, CO2, H2S and oxygen, and although they may be present in the furnace effluent in concentrations of less than 1%, they effectively bind up a portion of the sulphur which does not completely hydrolyze back to H2S in the catalyst beds of the sulphur plant, thus reducing the overall conversion of H2S to sulphur.
It is apparent that there is a clear need for a process that will increase the concentration of H2S in the feed gas entering a sulphur plant. Preferably the process should improve the conversion of H2S to sulphur, and should also solve many of the operational problems associated with feed gases that are too lean in H2S.
b) Relevant Technology
Advances toward improvement of H2S/CO2 ratios in sulphur plant feed have generally been based on the selectivity of methyldiethanol amine (MDEA) for H2S over CO2 when in contact with sour gas. Tertiary amines such as MDEA and also di-isopropyl amine (DIPA) exhibit this preferential affinity for H2S. Other amines such as MEA and DEA tend not to exhibit significant preferential affinity, and will therefore strongly absorb both H2S and CO2.
In studying the relative affinities between tertiary amines and the acid gases hydrogen sulphide and carbon dioxide, two characteristics must be considered. One is reaction equilibrium, which is defined as the final concentrations of reactants and reaction products after sufficient time has elapsed to attain steady levels. Equilibrium in thermodynamic terms occurs when the total free energy of the mixture reaches a minimum. The second characteristic to consider is reaction kinetics, which refers to the rate at which a reaction occurs. While consideration of reaction equilibrium is important, in the practical application of industrial chemistry, consideration of reaction kinetics is equally important since reaction time will greatly influence the final distribution of components in a reaction mixture. Such is the case with the tertiary amines, and also with DIPA. While the reaction with H2S is rapid, the reaction with CO2 is slow. Therefore, although consideration of reaction equilibrium alone would suggest that both H2S and CO2 could react almost to completion, when the reaction kinetics are considered, only the H2S reaction approaches completion, while the CO2 reaction goes only part way. Selective absorption of H2S can therefore be improved by limiting contact time. The mechanical design of contacting equipment, the operating conditions, and the presence of special chemical promoters can all have a bearing on selectivity of tertiary amines for H2S over CO2.
The popular amines MEA, DEA, MDEA, DGA, and DIPA all have in common a trivalent nitrogen atom to which are attached alcohol radicals (either ethanol or propanol). For example, the primary amine, monoethanol amine, has one ethanol group and two free hydrogen atoms. The secondary amine, diethanol amine, has two ethanol groups (as the name suggests) and one hydrogen atom. DGA has a single ether-ethanol chain with two hydrogens. MEA, DEA, and DGA all react rapidly with carbon dioxide, combining with the available proton of the amine molecule to form a carbamate radical (see FIG. 1). DIPA, which has two propanol structures and a single hydrogen atom, is not fully substituted, and is therefore not a tertiary amine. DIPA does not exhibit the rapid reaction with CO2 that is characteristic of the primary and secondary amines, each of which have an available proton. Apparently, the proton is not available for reaction with CO2, so the carbamate reaction does not occur readily with DIPA. Other amines, the so called hindered amines use the physical structure of the amine molecule to block accessibility of CO2 to the reactive hydrogen. Methyl diethanol amine (MDEA) is a tertiary amine which has no proton attached to the nitrogen atom. As the name suggests, the three valences of nitrogen are occupied by two ethanol groups and one methyl group, so the carbamate reaction, which requires a labile proton, cannot occur.
The reaction between a molecule of MDEA and a molecule of CO2 is somewhat more complex. When a CO2 molecule is dissolved in an aqueous solution, due to its acid nature it hydrolyzes to form carbonic acid (H2CO3). In a process which occurs slowly, the carbonic acid then dissociates to form positive hydrogen ions and negative bicarbonate ions. The bicarbonate may, to some extent, dissociate further to form additional positive hydrogen ions and negative carbonate ions. The MDEA molecule, being mildly basic in character, will bond loosely with the available hydrogen ions to form a positively charged amine-hydrogen ion that coexists in solution with negatively charged bicarbonate and carbonate ions (see FIG. 2). Since the carbonic acid dissociation step is relatively slow kinetically, the overall sequence of steps must also proceed slowly. The overall kinetic acid-base reaction between tertiary amines and carbon dioxide must therefore occur quite slowly. In contrast, the acid-base reaction of hydrogen sulphide occurs rapidly. In typical contacting devices, the H2S reaction rate is at least ten times faster than the CO2 reaction. These differential rates of reaction help to explain the selectivity of tertiary amines for H2S over CO2.
As the reaction between the amine and acid gas proceeds, more of the available amine molecules become bound to acid gas molecules, leaving fewer unreacted amine molecules available to react with the acid gas. This lack of available reactive amine molecules in the presence of acid gas slows the rate of reaction. Solution loading is therefore another factor influencing the selectivity of tertiary amines for H2S.
Reaction kinetics, however, is only one factor to consider in analyzing the absorption of acid gases by amine solutions. As in physical absorption, acid gas molecules must migrate to the gas liquid interface under the action of the concentration gradient that exists in the gas film adjacent to the interface. The molecule must then penetrate the interface and migrate inward until an unreacted amine molecule is encountered. As the mass transfer of acid gas molecules from the bulk gas phase into the liquid phase occurs by diffusion, the process of transfer requires a finite amount of time. Diffusion in this case occurs in two sequential steps. First, diffusion through the gas phase occurs near the interfacial boundary at the gas diffusion rate and, second, diffusion through the liquid phase occurs near the liquid boundary of the interface at the liquid diffusion rate. As a rate determining factor for tertiary amines, mass transfer by diffusion must be considered in addition to chemical rates of reaction. It has also been observed that selectivity for H2S increases as contact pressure decreases.
As previously mentioned, H2S reacts almost instantly with amine, so mass transfer by diffusion through the gas phase is the rate-limiting step for hydrogen sulphide. For carbon dioxide, the dissociation to form hydrogen and bicarbonate ions proceeds so slowly that the concentration gradient in the liquid phase that drives the mass transfer is impeded. This impedance constitutes an additional resistance to absorption of CO2.
For many reactions the temperature at which the reaction takes place has a profound effect on the rate at which the reaction occurs. This is particularly true of the reaction between CO2 and tertiary amines during absorption where a reduction in temperature of even a few degrees can significantly slow the rate at which the CO2 and amine react. The H2S reaction however is not as dramatically effected by a reduction in reaction temperatures, and this characteristic can be used to increase the differential rates of reaction between the two acid gases and thus increase the amines selectivity for H2S. By slowing the rate of the CO2 reaction by reducing the temperature in the absorber the “slipping” of CO2 can be increased without significantly affecting the absorption of H2S.
Obviously, one way to reduce the operating temperature of the absorber is to reduce the temperature of the streams feeding into the absorber. The reaction of acid gas with amine is exothermic however, and in some cases considerable amounts of heat can be generated by the reaction within the absorber itself. So even though the feed streams to the absorber are relatively cool, the heat of reaction will create a bulge in the temperature profile somewhere in the midsection of the tower. One way to mitigate this bulge is to use a more dilute amine solution so that the greater mass of the solution will attenuate the temperature rise. A more direct way to reduce the temperature bulge is to use inter-stage cooling on a stage near the peak of the temperature bulge. This will reduce the contact temperature throughout the tower which will in turn slow the CO2 reaction and will improve the amine's selectivity for H2S.
The extent of reaction between acid gas and amine is the product of the rate of reaction and the contact time of the reactants. Slowing the rate of reaction will improve the amines affinity for H2S over CO2. Reducing the contact time between the reactants will have a similar effect. The object should be to choose a type of contact device that can absorb H2S to meet the required specification in the briefest possible time so as to prevent the CO2 reaction from proceeding any further than necessary. The choice of an efficient multistage contacting device, whether it be conventional trays, random packing, structured packing or equivalent, can also have a significant effect on the selectivity of the amine for H2S.
Practical applications for the selectivity of tertiary amines for H2S over CO2 have, for the most part, been limited to absorption of acid gases from natural gas in a primary absorber (see FIG. 3 which shows a standard arrangement). Circulation rate and residence time in the absorber permit a portion of the CO2 to remain unabsorbed while H2S is totally removed from the gas. Commercial specifications for natural gas require near to total removal of H2S, but in most cases up to 2% carbon dioxide in the purified gas is acceptable. The tertiary amine, methyldiethanol amine (MDEA), is usually the preferred absorbent. The practice of partially removing CO2 from the natural gas is referred to as “slipping” the CO2.
In the technical record, references to MDEA's preferential affinity for H2S over CO2 appear as early as 1950, when Frazier and Kohl first noted the phenomena (see Is Frazier, H. D. and A. L. Kohl, “Selective Absorption of Hydrogen Sulfide from Gas Streams”, Ind. Eng. Chem., 42, 2258–2292 (1950)). Since then, the technical literature has traced the development of design methods for the use of MDEA. By the 1980's MDEA had gained widespread use in the gas industry, but applications were generally restricted to the relatively simple operation of slipping a portion of the CO2 in the high pressure absorber while totally absorbing the H2S. The formidable challenges of quantitatively predicting the combined chemical reaction and mass transfer relationships were not met until recent years, and although present methods are adequate, there is still significant room for improvement.
Present methods involve computational procedures to establish both chemical and mass transfer equilibrium relationships between the amine and the acid gases. The concentrations of the various chemical species seek to arrive at final equilibrium concentrations at which point no further change will occur. It is the difference between actual concentrations and equilibrium concentrations that provides the driving force for change to occur. Because there are various resistances to these changes, change does not to occur instantaneously; it occurs at a definite rate determined by the nature of the components, and by process conditions. Rate of change is proportional to driving force and inversely proportional to resistance, so if driving force and resistance can be calculated, the rate of change can also be calculated. If infinite time were available, equilibrium concentrations would eventually be attained. In reality, however, time constraints dictate that only a partial approach to equilibrium is attainable. This procedure forms the basis for the design of processing equipment to preferentially absorb H2S from gases containing a mixture of both H2S and CO2.
Since H2S proceeds toward equilibrium rapidly, it approaches equilibrium more closely than CO2, which proceeds slowly. In real absorbers, equilibrium can be approached, but is never attained. In a multistage contacting device such as a trayed tower, if each actual stage had sufficient time to reach equilibrium, the stages would be said to be 100% efficient. This hypothetical scenario provides a measure of the change that takes place on each actual stage if the actual change is expressed as a percentage of the change that would take place if equilibrium were attained. The actual change taking place on the stage could then be calculated from the known 100% efficiency of the stage if equilibrium is attained. For example, in a typical trayed MDEA absorber, the tray efficiency for H2S is approximately 50%, whereas the tray efficiency for CO2 is typically about one-tenth as much, or 5%. If this preferential effect is factored into multiple stages of contact, the separation of H2S from CO2 can be significant. In practical situations, however, it must be recognized that the final concentration of H2S in the treated gas must be very low, while the concentration of CO2 is many times higher. The driving force for absorption of H2S is low, while the driving force to absorb CO2 is relatively high in the top trays of the absorber tower. This means that, in the process of absorbing essentially all of the H2S, significant quantities of CO2 will inevitably also be absorbed, and that the rich MDEA exiting from the bottom of the absorber column will contain a large amount of CO2 along with the absorbed H2S.
Over the years various schemes have been proposed to improve the selectivity of tertiary amines for H2S over CO2, but unless the true complexity of the absorption process is recognized, the success of these schemes will be compromised. For example, many schemes attribute to the tertiary amines a strong similarity to physical absorption, in which acid gases are absorbed or desorbed in response to changes in pressure or temperature. Physical absorbents generally follow the principle of Henry's Law, which states that the concentration of a distributed component in the liquid phase is proportional to the partial pressure of the component in the gas phase. Due to chemical reactions that inevitably occur in the amine solution, amines do not behave in this manner. When the chemical bond between the amine and the acid gas is formed, it is not easily broken. Attempts to desorb the acid gases by pressure reduction, gentle heating, or gas stripping will therefore have only limited success. The only way to release significant amounts of acid gas from the amine solution is to break the chemical bond by vigorous steaming of the solution in the amine regenerator. Some proposed process schemes are based on mild partial regeneration to create a semi-lean amine solution, which because it is supposedly already loaded with CO2, will resist further absorption of CO2, and absorb H2S instead. Such schemes have not gained wide acceptance.